The occurrence rate of the earthquakes is determined in one of two ways. If the geologists can find evidence of distinct, datable ruptures in the past, the number of these ruptures is used to define a rate.
Otherwise, geologists estimate fault slip rate from accumulated scarp heights and estimated date for the oldest movement on the scarp. Given that a certain magnitude is likely to produce a displacement (slip) of a certain size during the earthquake, geophysicists can estimate the rate of occurrence of earthquakes of that magnitude.
Recurrence rates will be different for different assumed magnitudes thought to be “characteristic” of that fault. Generally, a smaller magnitude will produce a faster recurrence rate, and, hence, for moderate levels of ground motion, a higher hazard!
Quaternary faults are used to make fault sources for future earthquakes. Future earthquakes are assumed to be likely to occur where earthquakes have produced faults in the geologically recent past. The surface trace and downward extension of these faults is used to make a geometric model of the future fault plane.
How do we account for unknown active faults?
Two methods are used in cases in which there are faults known or suspected, but the faults have unknown slip rate.
In one case, there is evidence of seismic strain across a region, that is, the area is deforming in accordance with forces that cause faulting. However if this strain deformation is assigned only to the few known active faults in the region, the resulting slip rates are far larger than the known slip rates for those faults. It is then assumed that other faults must be accommodating the deformation. Accordingly, the strain is distributed uniformly across the region, assuming a range of magnitudes, a strike direction for the faults, and a b-value for the relative rates of occurrence of the various magnitudes. Then the hazard is calculated.
In the other case, there are areas where current seismicity is low or nil, but there is little other reason to believe that faulting is not likely to occur in the future. The historical seismicity is spread uniformly across areas of similar fault style. Again, magnitude ranges, strike directions and b-values are assumed. The areas are termed background zones. These background zones are used to generate hazard values on their own. At any location, if these background hazard values are greater than those hazard values calculated using the usual fault and seismicity sources, the background hazard values are used, instead. In effect, these background zones provide a floor of hazard, so that in inactive areas, with faults, there is never zero-valued hazard.